Although unemployment likely entails various externalities, research examining its spillover effects on spouses is scarce. This is the first paper to estimate effects of unemployment on the smoking behavior of both spouses. Using German Socio-Economic Panel data, we combine matching and difference-indifferences estimation, employing the post-double-selection method for control-variable selection via Lasso regressions. One spouse's unemployment increases both spouses' smoking probability and intensity. Smoking relapses and decreased smoking cessation drive the effects. Effects are stronger if the partner already smokes and if the male partner becomes unemployed. Of several mechanisms discussed, we identify smoking to cope with stress as relevant.
smoking risky health behaviors unemployment job loss spillover effects post-doubleselection method